Is a Spin-the-Wheel Really Random?

The animation is not the randomness

A real wheel in a casino or game show is random because of physics: friction, force and tiny variations no one can repeat. A wheel on a web page has none of that. It is a drawing, and its motion could be made to look any way a developer likes.

So the honest answer is that the wheel itself is not what makes the result random. What matters is the number generator that decides the winner, and whether that number is turned into a choice fairly. On this site the winner is chosen with crypto.getRandomValues(), and the wheel is the visual presentation of that result.

This is also why you should be wary of any claim that a wheel is “fair because it spins slowly” or “fair because it uses real physics”. Neither statement says anything about the code behind the page.

Pseudo-random versus true random

Computers rarely produce “true” randomness from physical noise on every request. Most use a pseudo-random generator: an algorithm that produces numbers that behave randomly, starting from a seed.

This is not a flaw for casual use. MDN describes Math.random() as pseudo-random with an approximately uniform distribution, suitable for things like games, but it adds that it does not provide cryptographically secure numbers. By contrast, the Web Crypto method getRandomValues() is designed for cryptographic use. MDN notes that it also relies on a pseudo-random generator, but one seeded with enough entropy to be suitable for cryptographic purposes.

For picking a winner, that means results are unpredictable in practice and not something a viewer could work out in advance. Entropy is simply unpredictability, and a generator seeded with plenty of it is hard to guess even for someone who knows the algorithm. That is the practical difference between casual and cryptographic randomness.

The hidden problem of modulo bias

Getting a random number is only half the job. The number must then be turned into a choice from your list. A common shortcut is to take the number and divide by the list size, keeping the remainder. This is slightly unfair.

A simple example shows why. A random byte has 256 possible values. With three entries, 256 does not divide evenly: one entry ends up with 86 values and the others with 85. The difference is tiny, but it is real, and it grows worse with larger ranges relative to the list.

The fix is to discard numbers that fall in the uneven leftover range and draw again. This tool does that, so every entry has the same chance. Most people never need to think about this, but it explains why careful tools use a slightly more involved method than the shortcut. The extra care costs nothing and removes a bias that would otherwise be invisible to the people watching.

Why streaks and repeats feel wrong

People expect random results to look evenly mixed. They do not. With ten entries, each has a one-in-ten chance on every spin, whatever happened before. The same entry can win twice in a row, and an entry can go many spins without winning.

Believing that a long-absent entry is “due” is a well-known mistake, often called the gambler’s fallacy. The wheel has no memory, and past spins do not change the next one.

This is why a handful of spins cannot show that a wheel is biased or fair. Judging that needs the method, not the outcomes. A useful mental check is to flip a coin. Getting five heads in a row is unusual but not suspicious, and over many flips, such runs appear regularly. Wheels behave the same way, and the more entries there are, the more varied the runs look.

Fair to the list, not fairer than the list

A wheel can only be as fair as the entries put in it. If one person appears twice, they have double the chance. The duplicate-removal option, which ignores capitalisation and is on by default, guards against accidental double entries.

The shuffle option, which is off by default, changes where entries sit on the wheel for each spin, and each entry keeps its colour. It makes the layout less predictable to watch, but it does not change anyone’s chance of winning, which is always one divided by the number of entries.

Who gets onto the list, and when entries close, are human decisions, and they affect fairness far more than the generator does. If you remove duplicates before the spin and keep the entry period transparent, the remaining fairness question is easy to answer: everyone on the list has an equal chance.

When a wheel is the right tool

A wheel is a good fit for classroom activities, team choices, streams and informal giveaways where participants can see the draw. It is a poor fit where a decision needs expertise, accountability or a legal process.

Nothing is uploaded or stored during a spin, which keeps things private, but it also means you must keep your own record if you need one. Rules for prize draws differ by country and platform, and nothing here is legal advice.

For fun and low-stakes picks, a wheel that uses a secure generator and avoids modulo bias is as random as you need it to be. If you ever need a draw that can be audited by a third party, a visible wheel on screen is not the same as a certified process. For regulated draws, follow the rules that apply to you instead.